Developers
June 20, 2020

OpenAI Launches an API that Allows Access to its NLP Models

A commercial API that focuses on the safe development of AI.
Source: Pixabay

Today we will talk about an API release by one of the biggest AI companies. OpenAI releases an API to give access to AI models.

This is not your regular API, most API´s are designed and built for one use-case. This API provides a text-in text-out interface. You can request access to the API so you can integrate it into your product. You can develop a new application or work on existing ones.  

You provide a prompt, and the API returns a text completion. You then program it so it does what you request. How do you achieve this? You show examples so that it understands what is it that you would like it to do. 

You can also train a dataset, it doesn't matter if it's big or small. The dataset is based on examples you provide and by learning from feedback from users.

The API stands in the middle of simple and powerful. It can be used by anyone, and at the same time, it works for large teams. It helps many companies to make Machine Learning more productive.

As AI progresses, there are every day more AI applications. In them, you can find good and bad applications. What does it mean good and bad applications? Well, not everybody has good intentions. Some people spend their time and energy on developing harmful applications. This is prohibited by OpenAI, and they terminate any API access to this type of applications and developers.  

The application currently launched in private beta, and not publicly available. This has one reason, and it is that in this way, OpenAI can manage better the content the API returns. They plan to launch small, but they surely dream big.

OpenAI counts with a big community focused on the development of positive AI systems. It is really interesting to see how an AI-focused company talks about the importance of developing positive applications.  

It's nothing new to mention, that AI has a lot of risk embedded. Teaching a computer to think like a human, and when tested seeing that it thinks even smarter, faster, and better than us can be something used for negative purposes too. It's a good practice to motivate developers to be conscious of what they are developing, as AI is a very powerful weapon.

Why did OpenAI release a commercial product?

When asked this question, OpenAI responds that what they mostly care about is about ensuring artificial intelligence benefits, everyone. The most obvious and logical reason of why they launched a commercial product, is to have revenue. This way they can have the necessary funding to grow and develop.

OpenAI decided to build an API instead of releasing an open-source model. There are a couple of reasons for this move. The first reason is the financial need. As we just described it, the company needs to find a way to survive economically.  

Secondly, delivering a documented and ready to be used API, allows much further reach than an open-source product that can only be implemented by big corporations (in this case). So, they intend to reach startups and small businesses by their API approach.

And lastly, it is much safer an API community than an open-source community. The API can be limited by closing doors to harmful application developers. In an open-source community, the negative developers could be reported and banned, but not in the easiness and effectiveness as in an API, where the owner just limits the usage and that's it.

How does OpenAI take action when they find harmful-biases?

 OpenAI acknowledges that it's hard not to fall into the harmful-bias in the AI industry. Their offering through an API tries to control this. The company says that they don’t even know if they are harming by their AI offer.

It is crystal clear that they are very consciously aware of the AI negative potential effects. They surely don't affect negatively, but they rather be cautious, so they can pick up any harmful-bias.

3 steps the company takes to help it be a safe community

  • Develop usage guidelines to help users develop positively and mitigate any risk
  • Work closely to developers to understand better their use cases and help them not fall into harmful-bias
  • Research about harmful-bias and report any new information to the community

In conclusion, OpenAI released an API to help developers work on their AI projects. The company makes a lot of emphasis on the importance of being aware of the dangers of AI. One of the most prominent interviews about the dangers of AI has been done to Elon Musk (founder of the company), who said that AI is even more dangerous than nuclear weapons.

The company seems to be following the same mindset, and part of their mission is to make AI progress most safely. The company launched a commercial API, so they could get revenue and they could still offer their services. They also claim that their commercial approach to AI is to be able to reach small startups and more businesses. OpenAI, one of the leaders in AI development, just released an API for their service and you might want to check it out.

TagsOpenAINLPAI
Lucas Bonder
Technical Writer
Lucas is an Entrepreneur, Web Developer, and Article Writer about Technology.

Related Articles

Back
DevelopersJune 20, 2020
OpenAI Launches an API that Allows Access to its NLP Models
A commercial API that focuses on the safe development of AI.

Today we will talk about an API release by one of the biggest AI companies. OpenAI releases an API to give access to AI models.

This is not your regular API, most API´s are designed and built for one use-case. This API provides a text-in text-out interface. You can request access to the API so you can integrate it into your product. You can develop a new application or work on existing ones.  

You provide a prompt, and the API returns a text completion. You then program it so it does what you request. How do you achieve this? You show examples so that it understands what is it that you would like it to do. 

You can also train a dataset, it doesn't matter if it's big or small. The dataset is based on examples you provide and by learning from feedback from users.

The API stands in the middle of simple and powerful. It can be used by anyone, and at the same time, it works for large teams. It helps many companies to make Machine Learning more productive.

As AI progresses, there are every day more AI applications. In them, you can find good and bad applications. What does it mean good and bad applications? Well, not everybody has good intentions. Some people spend their time and energy on developing harmful applications. This is prohibited by OpenAI, and they terminate any API access to this type of applications and developers.  

The application currently launched in private beta, and not publicly available. This has one reason, and it is that in this way, OpenAI can manage better the content the API returns. They plan to launch small, but they surely dream big.

OpenAI counts with a big community focused on the development of positive AI systems. It is really interesting to see how an AI-focused company talks about the importance of developing positive applications.  

It's nothing new to mention, that AI has a lot of risk embedded. Teaching a computer to think like a human, and when tested seeing that it thinks even smarter, faster, and better than us can be something used for negative purposes too. It's a good practice to motivate developers to be conscious of what they are developing, as AI is a very powerful weapon.

Why did OpenAI release a commercial product?

When asked this question, OpenAI responds that what they mostly care about is about ensuring artificial intelligence benefits, everyone. The most obvious and logical reason of why they launched a commercial product, is to have revenue. This way they can have the necessary funding to grow and develop.

OpenAI decided to build an API instead of releasing an open-source model. There are a couple of reasons for this move. The first reason is the financial need. As we just described it, the company needs to find a way to survive economically.  

Secondly, delivering a documented and ready to be used API, allows much further reach than an open-source product that can only be implemented by big corporations (in this case). So, they intend to reach startups and small businesses by their API approach.

And lastly, it is much safer an API community than an open-source community. The API can be limited by closing doors to harmful application developers. In an open-source community, the negative developers could be reported and banned, but not in the easiness and effectiveness as in an API, where the owner just limits the usage and that's it.

How does OpenAI take action when they find harmful-biases?

 OpenAI acknowledges that it's hard not to fall into the harmful-bias in the AI industry. Their offering through an API tries to control this. The company says that they don’t even know if they are harming by their AI offer.

It is crystal clear that they are very consciously aware of the AI negative potential effects. They surely don't affect negatively, but they rather be cautious, so they can pick up any harmful-bias.

3 steps the company takes to help it be a safe community

  • Develop usage guidelines to help users develop positively and mitigate any risk
  • Work closely to developers to understand better their use cases and help them not fall into harmful-bias
  • Research about harmful-bias and report any new information to the community

In conclusion, OpenAI released an API to help developers work on their AI projects. The company makes a lot of emphasis on the importance of being aware of the dangers of AI. One of the most prominent interviews about the dangers of AI has been done to Elon Musk (founder of the company), who said that AI is even more dangerous than nuclear weapons.

The company seems to be following the same mindset, and part of their mission is to make AI progress most safely. The company launched a commercial API, so they could get revenue and they could still offer their services. They also claim that their commercial approach to AI is to be able to reach small startups and more businesses. OpenAI, one of the leaders in AI development, just released an API for their service and you might want to check it out.

OpenAI
NLP
AI
About the author
Lucas Bonder -Technical Writer
Lucas is an Entrepreneur, Web Developer, and Article Writer about Technology.

Related Articles